Vaes et al. BMC Family Practice 2014, 15:113http://www.biomedcentral.com/1471-2296/15/113

RESEARCH ARTICLE Open Access

The diagnostic accuracy of the MyDiagnostick todetect atrial fibrillation in primary careBert Vaes1,2*, Silke Stalpaert1, Karen Tavernier1, Britt Thaels1, Daphne Lapeire1, Wilfried Mullens3,4

and Jan Degryse1,2

Abstract

Background: Atrial fibrillation is very common in people aged 65 or older. This condition increases the risk ofdeath, congestive heart failure and thromboembolic conditions. Many patients with atrial fibrillation areasymptomatic and a cerebrovascular accident (CVA) is often the first clinical presentation. Guidelines concerningthe prevention of CVA recommend monitoring the heart rate in patients aged 65 or older. Recently, theMyDiagnostick (Applied Biomedical Systems BV, Maastricht, The Netherlands) was introduced as a new screeningtool which might serve as an alternative for the less accurate pulse palpation. This study was designed to explorethe diagnostic accuracy of the MyDiagnostick for the detection of atrial fibrillation.

Methods: A phase II diagnostic accuracy study in a convenience sample of 191 subjects recruited in primary care.The majority of participants were patients with a known history of atrial fibrillation (n = 161). Readings of theMyDiagnostick were compared with electrocardiographic recordings. Sensitivity and specificity and their 95%confidence interval were calculated using 2x2 tables.

Results: A prevalence of 54% for an atrial fibrillation rhythm was found in the study population at the moment ofthe study. A combination of three measurements with the MyDiagnostick for each patient showed a sensitivity of94% (95% CI 87 – 98) and a specificity of 93% (95% CI 85 – 97).

Conclusion: The MyDiagnostick is an easy-to-use device that showed a good diagnostic accuracy with a highsensitivity and specificity for atrial fibrillation in a convenience sample in primary care. Future research is needed todetermine the place of the MyDiagnostick in possible screening or case-finding strategies for atrial fibrillation.

Keywords: Atrial fibrillation, Sensitivity and specificity, MyDiagnostick, Primary care

BackgroundAtrial fibrillation is currently the second most frequentlyoccurring cardiac arrhythmia in clinical practice (afterextrasystole). Its prevalence is estimated around 0.4 –1.0% in the general population, increasing with age, upto 5.0 – 6.0 in those aged 65 or older and 8.0% in sub-jects aged 80 or older [1].More than 30% of patients with atrial fibrillation are

asymptomatic. Furthermore, this condition increases therisk of death, congestive heart failure and thrombo-embolic conditions [2,3]. The risk of cerebrovascular

* Correspondence: Bert.Vaes@uclouvain.be1Department of General Practice, Katholieke Universiteit Leuven (KUL),Leuven, Belgium2Institute of Health and Society, Université Catholique de Louvain (UCL),Brussels, BelgiumFull list of author information is available at the end of the article

© 2014 Vaes et al.; licensee BioMed Central LtCommons Attribution License (http://creativecreproduction in any medium, provided the orDedication waiver (http://creativecommons.orunless otherwise stated.

accident (CVA) even increases fivefold in case of atrialfibrillation and CVA is often the first clinical presenta-tion of atrial fibrillation. Moreover, the risk for CVA in-creases with increasing age; from 9.9% in patientsbetween 70 – 79 years to 24% in the group of patientsbetween 80 – 89 years [4-6].The current European guidelines on atrial fibrillation

emphasize the importance to start anticoagulation inorder to decrease the risk of CVA [6]. In order to assessthis risk, the CHADS2 score is used, along with the morerecent and more accurate CHA2DS2VASc score [7,8].Therefore, the high prevalence of atrial fibrillation, thepotentially serious consequences of the condition, thepresence of effective therapy and the minimal impact oftreatment on the quality of life of the patient ensure that

d. This is an Open Access article distributed under the terms of the Creativeommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andiginal work is properly credited. The Creative Commons Public Domaing/publicdomain/zero/1.0/) applies to the data made available in this article,

Vaes et al. BMC Family Practice 2014, 15:113 Page 2 of 7http://www.biomedcentral.com/1471-2296/15/113

atrial fibrillation is a condition for which screening canor should be done [9].Guidelines concerning the prevention of CVA recom-

mend monitoring the heart rate in patients aged 65 orolder, followed by electrocardiogram (ECG) monitoringin case of an irregular rhythm [6,10]. Palpation of thepulse, to monitor the heart rate, has a good sensitivitybut a low specificity, making this a useful method of rul-ing out atrial fibrillation but a less good method of con-firming it [11]. Recently, the MyDiagnostick (AppliedBiomedical Systems BV, Maastricht, The Netherlands)[12] was introduced as a new screening tool and an al-ternative for the less accurate pulse palpation. This de-vice has the form of a rod with a metal handle on bothends. There are electrodes in these handles making itpossible to record a single-lead ECG (Figure 1). Thisnew screening technique would facilitate home screen-ing and increase the likelihood of detection of asymp-tomatic atrial fibrillation.Therefore, this study was designed to explore the diag-

nostic accuracy of the MyDiagnostick for the detectionof atrial fibrillation.

MethodsStudy populationParticipating general practitioners were asked to invitepatients with known, paroxysmal or chronic atrial fibril-lation to participate in the study. Furthermore, this con-venience sample was added up with subjects without ahistory of atrial fibrillation. With the probability of find-ing a false positive result of 5% or less (α = 0.05), an esti-mated prevalence of atrial fibrillation of 50% in thestudy population, an expected sensitivity and specificityof 95% and a confidence interval of 4%, a sample size of160 subjects was needed [13]. To end up with a preva-lence of atrial fibrillation of at least 50% at the momentof the study the vast majority of the invited patientswere people with a known history of atrial fibrillation(161/191) and only 30 people without a history of atrial

Figure 1 The MyDiagnostick device.

fibrillation. The patients who were willing to take part inthe survey were given an appointment for an examin-ation by the research team in their own general practicecentre. Before participating in the study, all patientssigned an informed consent. The study was approved bythe ethical review board of the Medical Faculty of theCatholic University of Leuven, Belgium (no ML 10464).Patients wearing a pacemaker were excluded if the pace-maker was configured in active pacing mode.Clinical characteristics were registered by the re-

searcher (SS, KT, BT or DL). The body mass index wascalculated, the blood pressure measured and a limitednumber of comorbidities extracted from the electronicmedical record. The presence of diabetes mellitus typeII, arterial hypertension and peripheral arterial illnessand a history of coronary heart disease and transient is-chaemic attack or CVA was reported and the CHA2DS2VASc score for each participant calculated. The intake ofplatelet aggregation inhibitors or anticoagulants (war-farin or new oral anticoagulants) was registered.

A trial fibrillationEach participant was tested with the MyDiagnostick(Applied Biomedical Systems BV, Maastricht, TheNetherlands) [12] by a single researcher who was notblinded for the medical history of the patient. This de-vice has the form of a rod with a metal handle on bothends. In these handles electrodes make it possible to rec-ord a single-lead ECG that is analyzed automatically.The patient was asked to grasp the device by both han-dles. After one minute the ECG lead was analysed andLED indicators gave a red or green signal that could beinterpreted as the presence or absence of atrial fibrilla-tion. Three consecutive recordings with the MyDiagno-stick with a 1 – 2 minute interval were done. Figure 2shows single-lead ECGs provided by the MyDiagnostick.The MyDiagnostick has a storage capacity of 140 record-ings of 50 – 70 seconds.The method of detection of AF in the MyDiagnostick

is based on the measurement of R-R interval irregularity.Prior AF detection, the acquired ECG (1 minute) is pre-processed and R-waves are detected. From all detectedR-wave annotations, R-R intervals are computed andused as an input for AF detection. The AF algorithm cal-culates an overall AF score based on a base rhythm-,periodicity- and variability score. The base rhythm scoreis based on a normal sinus rhythm state-machine chain-ing normal R-R intervals, including occasional prema-ture intervals and short runs of tachycardia. Creation oflong chains reflects a fit of the sinus rhythm state-model, lowering the probability of AF. The periodicityand variability scores are based on the R-R autocor-relation function. Periodicity of R-R interval patternswill generate multiple correlation peaks, whereas R-R

Sinus rhythm

Atrial fibrillation

Figure 2 Single-lead ECGs provided by the MyDiagnostick (sinus rhythm and atrial fibrillation).

Vaes et al. BMC Family Practice 2014, 15:113 Page 3 of 7http://www.biomedcentral.com/1471-2296/15/113

interval irregularity will lower correlation at only asmall shift. The overall AF score is obtained by linearcombination of all scores and compared to a thresh-old, producing a dichotomous result (AF/no AF).Afterwards a 12-lead electrocardiogram (ECG) (gold

standard) was carried out once by the same researcher.The ECGs were done using digital machines (HeartScreen80 G-L (MediSafe Ltd, Winchester, United Kingdom),Welch Allyn CP 50 (Welch Allyn, New York, USA),Nihon Kohden Cardiofax S 1250 (Nihon Kohden,Tokyo, Japan) or Schiller AT101 (Schiller AG, Baar,Switzerland)) and the data were immediately printed.The ECGs were analyzed off-line on the basis of theMinnesota Code Classification System for Electrocar-diographic Findings [14] (code 8-3-1) by an experiencedcardiologist (WM), blinded for the software interpret-ation and the results from the MyDiagnostick.The readings of the MyDiagnostick were compared with

the electrocardiographic recording. The overall three mea-surements on the MyDiagnostick were viewed for eachpatient. A green light three times was interpreted as anegative result and a red light three times as a positive re-sult. The non-uniform results of the MyDiagnostick wereinterpreted in favour of the most common outcome (i.e.2x red and 1x green was interpreted as a positive result,while 1x red and 2x green was interpreted as a negativeresult).

Statistical analysisSensitivity and specificity and their 95% confidenceinterval were calculated using 2x2 tables (MedCalc®11.6.0.0 (Mariakerke, Belgium)). The positive and nega-tive predictive values were then estimated based on an

expected prevalence of atrial fibrillation of 6% in thepopulation aged 65 or older.

ResultsA total of 191 subjects participated in this study ofwhich 91 (48%) women. The mean age of the partici-pants was 74.6 ± 9.7 years (range 50 – 99 years). In total30 participants without a history of atrial fibrillation (allin sinus rhythm on ECG) and 161 participants that wereknown to have paroxysmal or chronic atrial fibrillationwere included (Figure 3). Of these latter subjects, 103showed atrial fibrillation on their ECG at the moment ofthe study. Thus, a prevalence of 54% for an atrial fibrilla-tion rhythm was found in the study population. Table 1shows the patient characteristics according to the pres-ence of atrial fibrillation. A pacemaker was present in 17participants, in 10 patients the pacemaker was active atthe moment of the ECG recording and they were there-fore excluded from further analysis. The analysis wascontinued with 181 participants (prevalence of atrial fib-rillation, 53%).The MyDiagnostick showed a positive result in 96 par-

ticipants and a negative result in 85 subjects (Table 2).No unsuccessful attempts were reported. In the event ofa uniform result (n = 166, 92%) the diagnosis was correctin 96% of cases (seven misdiagnoses, four false positivesand three false negatives). A total of 15 participantsshowed non-uniform results (nine positive results, andsix negative results), in which 10 participants were inter-preted correctly (seven true positives and three true neg-atives) and five were misinterpreted. On the basis ofthese results sensitivity for the MyDiagnostick of 94%(95% CI 87–98) and a specificity of 93% (95% CI 85–97)was obtained. Based on an expected prevalence of 6% in

Total study populationn = 191

History of paroxysmal or chronic atrial fibrillation

n = 161

Atrial fibrillation at the moment of the study

n = 103

Sinus rhythm at the moment of the study

n = 88

No history of atrial fibrillation

n = 30

Active pacemaker

n = 7

No or inactive pacemaker

n = 96

Active pacemaker

n = 3

No or inactive pacemaker

n = 85

Figure 3 Flow chart of study participants.

Table 1 Characteristics of the study population (n = 191)

Atrial fibrillation present Atrial fibrillation absent

(n = 103) (n = 88)

Age, mean ± SD 77 ± 8 71 ± 11

Male gender, n (%) 57 (55.3) 43 (51.1)

BMI, mean ± SD 27.7 ± 6.9 25.9 ± 6.5

Systolic blood pressure, mean ± SD 130 ± 19 130 ± 18

Diastolic blood pressure, mean ± SD 75 ± 14 75 ± 10

CHA2DS2VASC score, median (IQR) 3 (2 – 4) 3 (2 – 3)

Comorbidity

Diabetes mellitus type II, n (%) 26 (25.2) 15 (17.0)

Arterial hypertension, n (%) 94 (91.3) 62 (70.5)

Coronary heart disease, n (%) 9 (8.7) 16 (18.2)

TIA or CVA, n (%) 17 (16.5) 4 (4 .5)

Peripheral arterial illness, n (%) 4 (3.9) 4 (4.5)

Anticoagulation

No anticoagulation, n (%) 0 (0) 22 (25.0)

Platelet aggregation inhibitor, n (%) 6 (5.8) 24 (27.3)

Warfarin, n (%) 69 (67.0) 30 (34.1)

New oral anticoagulants, n (%) 28 (27.2) 12 (13.6)

SD: standard deviation; IQR: inter-quartile range; TIA: transient ischaemic attack; CVA: cerebro-vascular accident.

Vaes et al. BMC Family Practice 2014, 15:113 Page 4 of 7http://www.biomedcentral.com/1471-2296/15/113

Table 2 2x2 table to calculate the diagnostic accuracy of the MyDiagnostick

Atrial fibrillation present Atrial fibrillation absent Total

MyDiagnostick positive result 90 6 96

MyDiagnostick negative result 6 79 85

Total 96 85 181

Vaes et al. BMC Family Practice 2014, 15:113 Page 5 of 7http://www.biomedcentral.com/1471-2296/15/113

the population aged 65 or older a positive predictivevalue of 45% (95% CI 24 – 68) and a negative predictivevalue of 99% (95% CI 97 – 100) were estimated.

DiscussionIn this study of a convenience sample in primary care,the MyDiagnostick was able to accurately set the diagno-sis of atrial fibrillation with a high sensitivity and specifi-city. The sensitivity measures the proportion of actualpositives that have been correctly identified and the spe-cificity measures the proportion of the actual negativesthat have been appropriately recognized as such. Thesetwo measures are closely related to the concepts of typeI (a test that shows a patient to have a disease when infact the patient does not have the disease) and type II(a test that shows a patient not to have a disease whenthe patient does have the disease) errors. A perfect diag-nostic test would be 100% sensitive (no false negatives)and 100% specific (no false positives). However, in a lowprevalence setting like primary care and for possiblescreening purposes the sensitivity of a test, the capacityto rule out the disease, could be considered as the mostimportant measure.The results for the MyDiagnostick are in line with the

diagnostic accuracy found for other screening methodsand devices. Pulse palpation to detect an irregular heart-beat gave a high sensitivity (87% – 97%) but a moderateto low specificity (71% – 81%) [11,15]. Although the spe-cificity could possibly be improved by continued pulseregistration, but at the cost of decreasing sensitivity. TheMicrolife BP monitor (Microlife Corporation, Taipei,Taiwan), a modified sphygmomanometer that recordspulse intervals, gave a sensitivity of 100% and a specifi-city of 90% (95% CI 87 – 92) when multiple measure-ments were carried out [16]. Another study that alsoused a modified sphygmomanometer showed a sensitiv-ity of 95% (95% CI 93 – 98) and specificity of 86% (95%CI 84 – 98) in a single measurement and a sensitivity of97% (95% CI 91 – 99) and specificity of 89% (95% CI 85 –92) on 3 measurements [17]. A finger probe that regis-tered a pulse wave pattern for 30 seconds gave a sensitivityof 100% and a specificity of 91% [18]. Two studies, similarto the current study, also used a device that registered asingle-lead ECG in which the interpretation was doneafterwards by a primary care physician [17,19]. Somervilleet al. found a sensitivity of 96% (95% CI 80 – 100) and a

specificity of 98% (95% CI 91 – 100) [17], and Mant et al.found a sensitivity of 85% (95% CI 79 – 91) and a specifi-city of 86% (95% CI 85 – 88) with the use of a thoraciclead and a sensitivity of 83% (95% CI 75 – 89) and a speci-ficity of 89% (95% CI 87 – 90) with the use of a peripherallead [19]. The latter sensitivities and specificities weremoderate primarily due to the varying interpretation ofECGs by primary care physicians. By contrast, the advan-tage of the MyDiagnostick is the absence of an interpret-ation bias.This study was a phase II diagnostic study investigat-

ing the accuracy of the MyDiagnostick in a conveniencesample of healthy people and people with a history ofatrial fibrillation. However, before a clinical recommen-dation on the use of the MyDiagnostick in every dayclinical practice can be given, a phase III diagnosticstudy has to be performed, determining the clinical con-sequences of introducing the MyDiagnostick through arandomised trial. The place of the MyDiagnostick inpossible future screening programs for atrial fibrillationtherefore remains to be determined. The use of theMyDiagnostick seems much more practical than electro-cardiographic monitoring at home. It involves no skinelectrodes or wires and the use of the device does notrequire any experience or medical knowledge. Further-more, using the MyDiagnostick is faster and cheaper.The MyDiagnostick also has an additional advantage incomparison with a blood pressure monitor that detectsan irregular pulse, since every recording is registeredand the single-lead ECGs can be consulted later on toconfirm the red lights that the patient detected.In terms of user friendliness, this screening tool is easily

manageable in the population aged 65 or older. The patientonly needs to hold the ends of the rod with both hands. Theunit goes on automatically, so that a visual limitation cannotbe a restricting factor. Furthermore, holding with the palmof the hand up or down, holding it lightly or firmly, trem-bling and physical activity during use do not seem to influ-ence the accuracy and uniformity of the results.In addition to detecting paroxysmal atrial fibrillation

and initiating the necessary treatment, the following ad-vantages of future use could be hypothesized: followingup on the response to electroconversion and medicaltreatment, improving compliance with therapy and po-tentially decreasing medical costs (by means of the pre-vention of ischemic CVA).

Vaes et al. BMC Family Practice 2014, 15:113 Page 6 of 7http://www.biomedcentral.com/1471-2296/15/113

Strengths and limitationsThis is the first study investigating the diagnostic accur-acy of the MyDiagnostick in primary care. High sensitiv-ities and specificities were obtained in the studypopulation, making this easy-to-use device a possiblecandidate to implement in future screening or case-finding programs for atrial fibrillation. However, a fewlimitations of the current study should be mentioned.First, a prevalence of atrial fibrillation of 6% was as-sumed in subjects aged 65 and older. However, a varietyof prevalence figures is found in literature and is mainlythe result of differences in the study population, themethods used (Holter, serial ECGs …) and the durationof ECG recording. Second, several ECG devices wereused as the gold standard instead of one standardizeddevice. Third, the possible place of the MyDiagnostick ina screening or case-finding strategy was not determinedwith the current study, nor was a cost-benefit analysis ofa screening or case-finding strategy with the MyDiagnostickdone. And fourth, because the study population was aconvenience sample extrapolation of these results tothe general population should be made with caution.Furthermore, investigations concerning the reliability ofthe MyDiagnostick for other arrhythmias and test-retestreliability must still be done.

ConclusionThe MyDiagnostick is an easy-to-use device that showeda good diagnostic accuracy with a high sensitivity andspecificity for atrial fibrillation in a convenience samplein primary care. Future research is needed to determinethe place of the MyDiagnostick in possible screening orcase-finding strategies for atrial fibrillation.

AbbreviationsCVA: Cerebrovascular accident; CI: Confidence interval;ECG: Electrocardiogram; SD: Standard deviation; IQR: Inter-quartile range;TIA: Transient ischaemic attack.

Competing interestsThe author declares that they have no competing interests.

Authors’ contributionsBV, SS, KT, BT, DL drafted the manuscript. JD is responsible for the design,conduct and analysis of the study. All authors participated in the criticalrevision of the manuscript. All authors read and approved the finalmanuscript.

AcknowledgementsThe authors wish to thank the following general practitioners for contactingpatients and providing consultation rooms: Dr. K. Van der Veken, Dr. C. VanCauwenberge, Dr. K. Schelfaut, Dr. N. Van Den Haute, Dr. J. Cooreman, Dr. L.Bontinck, Dr. R. Mechiels, Dr. K. Verhaegen, Dr. Z. De Aguirre, Dr. A. Van Hyfte,Dr. F. Froyman, Dr. V. Heylen, Dr. L. Termote, Dr. M. Vandenheede, Dr. S.Vandeputte, Dr. A. Verraes, Dr. K. Van Boeckel, Dr. L. Bruyninckx, Dr. L. Froyen,Dr. B. Schoolmeesters, Dr. R. Schoolmeesters, Dr. G. Snijkers, Dr. K. Van Dael,Dr. J. Gryffroy, Dr. L. Cras, Dr. F. Martens, Dr. C. Tack, Dr. L. Vanneste. Wewould also like to thank the firm Applied Biomedical Systems BV for theprovision of the MyDiagnosticks.

Author details1Department of General Practice, Katholieke Universiteit Leuven (KUL),Leuven, Belgium. 2Institute of Health and Society, Université Catholique deLouvain (UCL), Brussels, Belgium. 3Department of Cardiology, Ziekenhuis OostLimburg, Genk, Belgium. 4Biomedical Research Institute, Faculty of Medicineand Life Sciences, Hasselt University, Diepenbeek, Belgium.

Received: 1 February 2014 Accepted: 5 June 2014Published: 9 June 2014

References1. Pérula-de-Torres LA, Martínez-Adell MA, González-Blanco M, Baena-Díez JM,

Martín-Rioboó E, Parras-Rejano JM, González-Lama J, Martín-Alvarez R,Ruiz-Moral R, Fernández-García JÁ, Pérez-Díaz M, Ruiz-de-Castroviejo J,Pérula-de-Torres C, Valero-Martín A, Roldán-Villalobos A, Criado-Larumbe M,Burdoy-Joaquín E, Coma-Solé M, Cervera-León M, Cuixart-Costa L,Collaborative Group DOFA-AP: Opportunistic detection of atrialfibrillation in subjects aged 65 years or older in primary care: arandomised clinical trial of efficacy. DOFA-AP study protocol.BMC Fam Pract 2012, 13:13–106.

2. Miyasake Y, Barnes ME, Gersh BJ, Cha SS, Bailey KR, Seward JB, Iwasaka T,Tsang TS: Coronary ischemic events after first atrial fibrillation: risk andsurvival. Am J Med 2007, 120:357–363.

3. Wang TJ, Larson MG, Levy D, Vasan RS, Leip EP, Wolf PA, D'Agostino RB,Murabito JM, Kannel WB, Benjamin EJ: Temporal relations of atrialfibrillation and congestive heart failure and their joint influenceon mortality: the Framingham Heart Study. Circulation 2003,107:2920–2925.

4. Morgan S, Mant D: Randomised trial of two approaches to screeningfor atrial fibrillation in UK general practice. Br J Gen Pract 2002,373:380.

5. Markides V, Schilling RJ: Atrial fibrillation: classification, pathophysiology,mechanisms and drug treatment. Heart 2003, 89:939–943.

6. Camm J, Lip G, De Caterina R, Savelieva I, Atar D, Hohnloser SH, Hindricks G,Kirchhof P, ESC Committee for Practice Guidelines(CPG): 2012 focusedupdate of the ESC Guidelines for the management of atrial fibrillation.Eur Heart J 2012, 33:2719–2747.

7. Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford MJ:Validation of clinical classification schemes for predicting stroke:results from the National Registry of Atrial Fibrillation. JAMA 2001,285:2864–2870.

8. Lip G, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ: Refining clinical riskstratification for predicting stroke and thromboembolism in atrialfibrillation using a novel risk facto-based approach: the Euro HeartSurvey on Atrial Fibrillation. Chest 2010, 137:263–272.

9. Wilson JMG, Jungner G: The principles and practice of screening for disease.WHO Public Health papers. Geneva: World Health Organization; 1968.

10. Nieuwlaat R, Olsson SB, Lip G, Camm AJ, Breithardt G, Capucci A,Meeder JG, Prins MH, Lévy S, Crijns HJ, Euro Heart Survey Investigators:Guideline-adherent antithrombotic treatment is associated withimproved outcomes compared with undertreatment in high-risk patientswith atrial fibrillation: the Euro heart Survey on Atrial Fibrillation.Am Heart J 2007, 153:1006–1012.

11. Harris K, Edwards D, Mant J: How can we best detect atrial fibrillation?J R Coll Physicians Edinb 2012, 42(suppl 18):5–22.

12. MyDiagnostick: https://www.mydiagnostick.com/home-en.13. Carley S, Dosman S, Jones SR, Harrison M: Simple nomograms to calculate

sample size in diagnostic studies. Emerg Med J 2005, 22:180–181.14. Prineas R, Crow R, Blackburn H: The Minnesota Code Manual of

Electrocardiographic Findings. John Wright – PSG, Inc: Littleton MA; 1982.15. Wiesel J, Abraham S, Messineo FC: Screening for asymptomatic atrial

fibrillation while monitoring the blood pressure at home: trial of regularversus irregular pulse for prevention of stroke (TRIPPS 2.0). Am J Cardiol 2013,111:1598–1601.

16. Wiesel J, Fitzig L, Herschman Y, Messineo FC: Detection of atrial fibrillationusing a modified microlife blood pressure monitor. Am J Hypertens 2009,22:848–852.

17. Somerville S, Somerville J, Croft P, Lewis M: Atrial fibrillation: a comparisonof methods to identify cases in general practice. Br J Gen Pract 2000,50:727–729.

Vaes et al. BMC Family Practice 2014, 15:113 Page 7 of 7http://www.biomedcentral.com/1471-2296/15/113

18. Lewis M, Parker D, Weston C, Bowes M: Screening for atrial fibrillation:sensitivity and specificity of a new methodology. Br J Gen Pract 2011,61:38–39.

19. Mant J, Fitzmaurice DA, Hobbs FD, Jowett S, Murray ET, Holder R, Davies M,Lip GY: Accuracy of diagnosing atrial fibrillation on electrocardiogram byprimary care practitioners and interpretative diagnostic software: analysisof data from screening for atrial fibrillation in the elderly (SAFE) trial.BMJ 2007, 335:380.

doi:10.1186/1471-2296-15-113Cite this article as: Vaes et al.: The diagnostic accuracy of theMyDiagnostick to detect atrial fibrillation in primary care. BMC FamilyPractice 2014 15:113.

Submit your next manuscript to BioMed Centraland take full advantage of:

• Convenient online submission

• Thorough peer review

• No space constraints or color figure charges

• Immediate publication on acceptance

• Inclusion in PubMed, CAS, Scopus and Google Scholar

• Research which is freely available for redistribution

Submit your manuscript at www.biomedcentral.com/submit